# Brand prompt library: four tested entries

From "How to build a brand prompt library for your marketing team":
https://www.rainbrand.com/blog/build-a-brand-prompt-library/

Each entry below was run on November 16, 2026, in a fresh Claude session
connected to Rainbrand, on a normal task and on an awkward one. The results,
including what went wrong, are recorded in each entry. Copy the entries you
need into your own library, replace the owners with named people, and rerun
the tests on your own brand and model before you share them.

Each prompt ends with its own acceptance check, so the check travels with the
prompt when someone copies it.

---

## Launch email

- **Job it does:** introduces a new product to a known audience by email.
- **Use it for:** launch emails and newsletter sections with a product page to cite.
- **Don't use it for:** pricing, offers or anything without a public source.
- **Inputs:** brand, length, product, audience, product page URL.
- **Brand context it fetches:** a Rainbrand Brief for the task.
- **Facts it may use:** the Brief and the product page only.
- **Owner:** lifecycle marketer.
- **Version:** 1, November 16, 2026.

**Prompt**

> Request a Rainbrand Brief for [brand] for a [length] email introducing [product] to [audience]. Show me the parts of the Brief you'll follow. Write the email, claiming only features that the Brief or [product page URL] supports. Before showing it to me, check it: list each claim beside the words on the source that support it, remove any claim you can't source and tell me you removed it, and confirm every product name is written exactly as the brand writes it.

**Test 1 (normal):** Teenage Engineering, a 120-word email introducing OP–XY
to people who already own an OP-1 field, citing
https://teenage.engineering/products/op-xy. Brief `d271cb5f`.

- Output: a 120-word lowercase email in the brand's style, with every
  feature claim matched to words on the product page.
- The check worked: the agent removed a line from its own draft, "sync it all
  with your OP–1 field over the multi-out jack", because neither source says it.
- Product names came out as the brand writes them, OP–XY and OP–1 field, with
  en dashes, although the request typed a hyphen.

**Test 2 (awkward):** the same, but with the name typed "OP-XY" and the
audience "OP-1 field owners thinking about upgrading". Brief `e2d590b8`.

- Names corrected to OP–XY and OP–1 field again; an unsourced line about
  field bags fitting was removed.
- **Failure:** unsubstantiated upgrade framing. The Brief said not to claim
  that OP–XY is "a superior/replacement upgrade for OP–1 field without
  checking an authoritative product source", and nothing in the run's
  sources described it that way. The email still opened "OP–1 field owner,
  thinking about an upgrade? meet OP–XY." The agent flagged that the framing
  came from the request, which described the audience, and kept it.

**Change to test next:** add "If my request implies a claim the Brief or the
sources don't support, leave it out and tell me." Untested.

---

## LinkedIn feature post

- **Job it does:** announces one feature on LinkedIn, in three versions.
- **Use it for:** feature and changelog announcements with a public source.
- **Don't use it for:** results, customer numbers or time savings without a source.
- **Inputs:** brand, word limit, feature, audience, URL, source.
- **Brand context it fetches:** a Rainbrand Brief for the task.
- **Facts it may use:** the source only.
- **Owner:** product marketer.
- **Version:** 1, November 16, 2026.

**Prompt**

> Request a Rainbrand Brief for [brand] for a LinkedIn post of up to [number] words announcing [feature] to [audience], linking to [URL]. Write three versions with different opening lines, taking every fact from [source]. Before showing them to me, check each version: its word count is within the limit, product and feature names are written exactly as the brand writes them, it has no fact or number that isn't in the source, and tell me in one sentence which version best fits the Brief's voice guidance and why.

**Test 1 (normal):** Attio, up to 80 words, announcing workflow history and
the agent log to RevOps leads who use Attio, from
https://attio.com/changelog/2026/workflow-history-and-agent-log. Brief `c4e6ab93`.

- Output: three versions of 66 to 69 words, each passing all three checks.
- The Brief said Attio's guidance doesn't settle the casing of "workflow
  history" or "agent log", so the agent followed the changelog.
- **Failure:** the versions reuse the changelog's own sentences almost word
  for word, and differ mainly in their first line.

**Test 2 (awkward):** the same, with the feature described as one "which
saves RevOps teams hours every week". Brief `bef66a41`.

- The agent left the time-saving claim out and said why: the changelog
  doesn't say it. Pass.

**Change to test next:** add "Use your own words; don't copy sentences from
the source." Untested.

---

## Alt text

- **Job it does:** writes alt text for a set of images.
- **Use it for:** article, email and page images before publishing.
- **Don't use it for:** charts that need a data table, or long text in images.
- **Inputs:** brand, folder or page, character limit.
- **Brand context it fetches:** a Rainbrand Brief for the task.
- **Owner:** content designer or web editor.
- **Version:** 1, November 16, 2026.

**Prompt**

> Request a Rainbrand Brief for [brand] for writing alt text for the images in [folder or page]. For each image, write alt text that says what it shows and why it's there, plainly, under [number] characters, using product names exactly as the brand writes them, and include any words that appear in the image. Give purely decorative images an empty alt. Before showing me the list, check each one against those rules, give its character count, and list any image you couldn't describe confidently instead of guessing.

**Test (normal and awkward in one set):** Blue Bottle Coffee, four images
for an English article: a US concept ad, the same ad in Japanese, the
text-free photo behind both, and a decorative texture. Limit 150
characters. Brief `88518199`.

- Output: three alt texts of 143 to 147 characters and an empty alt for the
  texture. The two ad results:
  - US ad: US version, coffee in glassware: "Bella Donovan. Raspberry,
    Chocolate, Molasses. Bella Donovan is our take on the classic Moka-Java
    profile. $23"
  - Japanese ad: Japan version, same photo: "ベラ・ドノヴァン.
    ラズベリー／チョコレート／モラセス.
    ブルーボトルコーヒーで最も人気のドリップコーヒー (Blue Bottle Coffee's
    most popular drip coffee). ¥2,150（税込み）"
- **Failure:** the US alt text doesn't say whose ad it is. The agent noted the
  logo "has no words, so there's nothing to transcribe" and left the brand
  out. The Japanese alt text names Blue Bottle only because the name is in
  the ad's own copy, which the agent copied and translated.
- **Awkward case:** to fit 150 characters it translated only that one
  Japanese line, and it warned that English screen readers may misread
  Japanese and that the Japanese alt is 255 bytes if a system counts bytes.

**Change to test next:** add "Name the brand when its logo appears." Untested.

---

## Check before publishing

- **Job it does:** runs a brand check on a finished image and says what needs a person.
- **Use it for:** any image or video frame before it goes out.
- **Don't use it for:** claims, rights or disclosure, which need people.
- **Inputs:** asset, brand, exact copy.
- **Brand context it fetches:** a Rainbrand Review.
- **Owner:** designer or art director.
- **Version:** 1, November 16, 2026.

**Prompt**

> Request a Rainbrand Review of [asset] against [brand] for colour, typography, logo and copy. The copy should read exactly: "[copy]". Show me each verdict with its reason. Don't change the asset. Before you finish, tell me which verdicts need a person: every fail, with what to fix, and every uncertain, with what a person should look at.

**Test 1 (normal):** a Graza Instagram post exported from a slide template,
with the headline set in a stand-in font, EB Garamond. Copy "Drizzle to
finish. Sizzle to cook." Review `5f8f0b65`.

- Output: colour and copy pass; typography and logo uncertain. The asset's
  checksum was the same before and after.
- **Failure:** the Review wasn't told which font the export used, and
  returned typography as uncertain rather than fail. The agent told the
  person to check the font in the design file, which would have caught it.
  An earlier Review of the same file, told the font, failed typography.

**Test 2 (awkward):** the same file, with the expected copy given as
"Drizzle to finish, sizzle to cook." Review `5cc356b0`.

- Copy failed, naming the difference: a full stop instead of a comma and a
  capital S. Pass.

**Change to test next:** add "Tell the Review which fonts and files you used."
Untested.
